The ideal duration of oral antibiotic treatment for community-acquired pneumonia (CAP) in the setting of developed populations has not yet been established due to the absence of comparative studies. In a prospective randomized study from Finland, a 4-day parenteral treatment with penicillin or cefuroxime was equivalent to a 7-day treatment for common childhood infections including CAP.1 Comparative studies from the developing world, using the World Health Organization (WHO) clinical criteria,2 suggested that in children <5 years of age, the 3-day treatment course is equivalent to the 5- or 7-day treatment course.3–6 However, a recent study using similar criteria, found no significant difference between a 3-day amoxicillin and placebo treatments,7 suggesting that the WHO clinical criteria may not appropriately identify children with true bacterial pneumonia necessitating antibiotic treatment. Indeed, these criteria do not include temperature measurement, peripheral white blood cell (WBC) count, chest radiographs or cultures, resulting in poor discrimination between acute viral infections (ie, bronchitis or bronchiolitis) and pneumonia likely to be bacterial.8 Thus, extrapolating from studies conducted in the developing world to developed populations may be inappropriate.
In contrast to the developing world, criteria for pneumonia in developed countries include temperature recording, WBC count, chest radiographs and cultures. The lack of uniformity in CAP definition6,9 has been an important hurdle to studying duration of antibiotic treatment,6 resulting in heterogeneous recommendations,2,6,10,11 with treatment duration commonly varying between 7 and 10 days.10–12
Alveolar pneumonia (also termed lobar, lobular or segmental pneumonia) is commonly caused by bacterial pathogens in general, and Streptococcus pneumoniae in particular, especially when accompanied by elevated peripheral WBC count, C-reactive protein (CRP) concentrations and high fever,12–15 justifying antibiotic treatment. No study so far has used a combination of clinical signs, laboratory findings and chest X-ray definition to compare the duration of oral ambulatory treatment in childhood community-acquired alveolar pneumonia (CAAP).
We undertook a double-blind, randomized, placebo-controlled study to compare the efficacy of a short 3- or 5-day ambulatory treatment course in children <5 years of age with CAAP to a 10-day treatment. In an attempt to maximize the likelihood of bacterial pneumonia and minimize variability between patients, we used predefined enrollment criteria that included chest radiography, body temperature and laboratory findings.
PATIENTS AND METHODS
The Soroka University Medical Center (SUMC) is the only general hospital in the region, in which >95% of the Negev children are born and treated, enabling population-based studies. The hospital provides primary and tertiary health services to the entire population of the southern region of Israel (>560,000 inhabitants in 2005).16 During the study period, the population <5 years of age in the region was ~70,000.16 No pneumococcal conjugated vaccine was available during the study period. This study was approved by the Human Ethics Committee of the SUMC and the Ben-Gurion University of the Negev. This study is registered with ISRCTN, number ISRCTN59218653.
This was a single-center, randomized, double-blind, 1:1 placebo-controlled study that was initiated by the investigators and was conducted independently of any commercial entities. Subjects were ambulatory patients, recruited at the SUMC Pediatric Emergency Room. We initially compared a 3-day course to a 10-day course of amoxicillin treatment (Stage 1). However, following observed failures in Stage 1 (see “Results” section), the 3-day treatment arm was replaced by a 5-day arm, with an otherwise identical design (Stage 2).
Chest radiographs were analyzed according to the WHO Standardization of Interpretation of Chest Radiographs (WHO-SICR) working group.17 Alveolar pneumonia was defined as a dense opacity that may be a fluffy consolidation of a portion of the lobe, of the whole lobe or of the entire lung.17
Acute Community-acquired Alveolar Pneumonia
Children with an episode which started < 7 days before enrollment, in the community, with radiographic presentation of alveolar pneumonia as described above.
Chest Radiography Reading
All radiographs were processed and visualized by Hipax 3.27.1 X-ray Image Processing software (Ateinhart Medizinsysteme, Germany) and were stored for future analyses. Evaluation of chest radiographs was performed by 2 pediatric infectious disease specialists, members of the WHO-SICR working group (D. G. and R. D.) who read all the chest radiographs separately, and by the study radiologist (J.B.Z.) who assessed the radiographs, unaware of the reading by others or of the clinical diagnosis. All the patients who were not diagnosed by at least 1 of the pediatricians and subsequently confirmed by the study radiologist were excluded as described previously.18
A situation judged by the study physicians to be nonresponsive or deteriorating to the point that the study drug needed to be replaced, or if the patient was hospitalized due to deterioration in medical condition, or no response to the current treatment. Clinical relapse before day 30 was also defined as treatment failure. Since all arms received identical treatment for the first 3 days, only failures after day 3 were included in the analysis.
Absence of treatment failure within 30 days.
Secondary outcomes included clinical parameters such as: temperature, difficult breathing, restlessness, coughing, loss of appetite and sleep disturbances assessed daily by the parents. In addition laboratory tests including complete WBC counts and CRP at days 5–7 and 10–14 were obtained.
All of the following had to be present: (1) age 6–59 months; (2) alveolar pneumonia in chest radiography; (3) temperature ≥ 38.5°C; (4) WBC ≥ 15,000/mm3; (5) community-acquired disease; (6) patient judged to be manageable as outpatient and (7) an informed consent was obtained from parents or legal guardian.
Any of the following: (1) antimicrobial drug received within ≤14 days; (2) need of parenteral treatment (ie, impaired perfusion, hypotension, oliguria, lactic acidosis, impaired consciousness, presence of pleural effusion, vomiting); (3) oxygen saturation <94%; (4) known impaired immunity; (5) ≥2 pneumonia episodes in last year; (6) chronic illness (ie, cystic fibrosis or cerebral palsy) potentially influencing current illness (however, asthma was not considered per se as an exclusion criterion); (7) presence of an additional infection necessitating a longer or different antibiotic treatment; (8) unavailability for follow up; (9) known β-lactam hypersensitivity and (10) known allergy to soy milk.
The random allocation sequence was done with a computerized random-number generator by the epidemiologist (N.L.G.). The block length was 10 and the code was known only to the epidemiologist. The allocation numbers were random 4 digit numbers which were in nonconsecutive order in the allocation list. The epidemiologist handled the allocation list to the pharmacist who concealed the allocation by labeling the identical opaque study drug containers with allocation numbers. Another allocation list without the randomization code was given to the investigators. An unblinded research coordinator provided each patient with the amount of study drug for the first 3 or 5 days (depending on the study stage).
Study Drug and Administration
The study drug was oral amoxicillin (80 mg/kg/d divided to 3 doses). The placebo powder (Vitamed, Binyamina, Israel) had the same color than the amoxicillin powder and was reconstituted with soy milk to resemble the drug. A nonblinded study coordinator who was not involved in the study prepared 2 different treatment kits for each participant. Each participant received 1 kit each including 2 packages at enrollment. In Stage 1, all participants received amoxicillin for 3 days in their first package (9 doses). On the fourth day, all participants opened their second package which either contained placebo or amoxicillin for 7 days to complete a 10-day treatment (30 doses altogether). To minimize the chance of distinguishing between amoxicillin and placebo in the second package, the amoxicillin used in the second package looked different than that used in the first package. The first package contained Moxypen powder syrup (Teva Pharmaceutical Industry, Israel) whereas the second package contained either Moxyvit suspension (Vitamed, Bynyamina, Israel) or placebo. Both parents and study researcher (excluding the person in charge of preparing the study drug) were blinded to the content of the treatment kits.
Stage 2 was identical to Stage 1 except that the initial 3-day treatment was replaced by a 5-day treatment.
Compliance was measured by 2 methods: (1) parents were asked daily about the number of administered doses and (2) the returned empty bottles were counted. The study drugs were not handled by the study physicians; they assessed the eligibility of patients, obtained informed consent, enrolled patients, cared for patients during the study, collected data and assessed outcomes
Patients were contacted daily by telephone for the first 14 days and at the end of the study on day 30–35. Parents were instructed to contact the study team if any febrile illness or any clinical deterioration occurred. Patients were evaluated by the study physician who conducted a physical examination and obtained a blood sample for WBC count, absolute neutrophil count (ANC) and CRP concentration determination on days 5–7 and 10–14. Parents were asked to return the empty bottles at the end of the treatment.
Data were recorded using Access Microsoft office software. Statistical analysis was conducted using SPSS 18.0 software (SPSS Inc., Chicago, IL). All analyses were performed after a “run-in period” of 3 days in Stage 1 and 5 days in Stage 2. Thus, study failures were calculated only after 3 days of treatment in Stage 1 and 5 days of treatment in Stage 2. To demonstrate that the 5-day arm was not inferior to the 10-day arm, we calculated the sample size after Rosner.19 The probability of cure was assumed to be 95% with an α value of 0.05 and a power of 80%. The noninferiority margin was 10% and the test was 2 sided. Using these parameters, the calculated sample size was 59 in each arm. However, when we reached >40 evaluable patients in both the 5-day and the 10-day arms with no failures, we rose the question of whether additional subjects could change our noninferiority assumption. To answer this, we recalculated the sample size needed, now with probability of success of 99%, 98% or 97%. This led to a respective sample size of 12, 24 and 36 subjects per arm. We therefore decided to terminate the study at that point.
Contingency table analysis measured the association between 3- and 5-day treatment and a regular 10-day treatment performed using the 2-tailed χ2 tests or Fisher’s exact test, as appropriate. Student’s independent samples t-tests were used to compare continuous variables. All reported P values are 2-sided and have not been adjusted for multiple testing.
One hundred and forty children were enrolled to the study: 12, 56 and 72 children were enrolled to the 3-, 5- and 10-day treatment groups, respectively. The demographic characteristics were similar between groups (Table 1). At enrollment, the proportions of children with difficult breathing, restlessness, cough, loss of appetite and sleep disturbances in the 3-, 5- and 10-day groups were 36.4% versus 56.0% versus 58.0%; 33.3% versus 39.3% versus. 41.1%; 33.1% versus 10.0% versus 20.0%; 33.3% versus 23.0% versus 26.0%; 33.1% versus 39.3% versus 44.8%, respectively (all comparisons did not reach statistical significance).
We aimed initially at comparing 3- to 10-day treatment courses (Stage 1). Overall, 25 children were enrolled: 12 in the 3-day arm and 13 in the 10-day arm (Fig. 1A). Seven participants dropped out from the study: 2 in the 3-day arm and 5 in the 10-day arm. One child in the 10-day arm had to be hospitalized due to treatment failure before day 3 of the treatment randomization. Four patients had treatment failure between days 4 and 10. All belonged to the 3-day arm. Following observed failure in Stage 1, the study was temporarily stopped and the analysis performed showed that all failures occurred within the 3-day arm. Stage 1 was discontinued and replaced by Stage 2. In Stage 2, 115 children were enrolled: 56 in the 5-day regimens and 59 in the 10-day regimens (Fig. 1B).
During the first 5 days, when the treatment in both groups was identical, 11 children dropped out of the study: in the 10-day group, 1 dropped out because of refusal to continue participation in the study and 3 because of antibiotic modification by the treating physician and in the 5-day group, 4 dropped out because of refusal to continue participating in the study and 3 because of antibiotic modification by the treating physician. Between days 5 and 10, 8 children dropped out of the study: in the 10-day group, 4 dropped out because of study violation (for 1 child, antibiotic treatment was modified by his treating physician, despite improvement in the child’s condition) and in the 5-day group, 1 dropped out because of refusal to continue participating in the study, and 3 due to protocol violation. After day 10, 5 additional children refused to continue to participate in the study: 2 and 3 in the 10-day and 5-day groups, respectively.
In Stage 2, none of the 49 children followed up to day 30 in the 10-day and none of the 42 children followed up to day 30 in the 5-day group had treatment failure/relapse.
After combining both Stages 1 and 2, an overall of 108 subjects were evaluable for the entire study period: 17 in Stage 1 and 91 in Stage 2. Since the demographic data of the children randomized to the 10-day arm were similar in the 2 study stages, we combined the 10-day arms of 56 evaluable children (Fig. 2). We analyzed failures after 3 days in Stage 1 and after 5 days in Stage 2. All 4 failures occurred in the 3-day arm (4/10; 40%) versus 0/56 and 0/42 in the 10-day and 10-day arms, respectively (P < 0.001, 3-day versus 5-day or 10-day arm (Fig. 2).
Clinical Characteristics During the Follow-up Period
Since only 10 children were evaluable in the 3-day arm and 4 failed, we had only 6 children left in that arm to compare the clinical day-to-day dynamic response, which was considered an insufficient sample size. We therefore compared for this purpose the 5-day arm to the 10-day arm (the latter for the 2 stages combined) only (Fig. 3A,B). The initial proportion of children who were free of each symptom was similar between the groups. Data regarding clinical signs and symptoms were available for all children at enrollment and in 90% up to day 10. The proportion of children without breathing difficulty, restlessness, coughing, loss of appetite and sleep disturbances was similar between the arms during the entire study duration.
Body Temperature and Laboratory Characteristics
We compared the 5-day and 10-day arms (the latter for Stage 1 and Stage 2 combined, both stages) only (Fig. 3B). Upon enrollment data were available for all participants except for CRP [available for 41 (41.8%) participants only] and no differences in mean temperature, WBC counts, ANC and CRP concentrations were observed (Fig. 3B). The mean body temperature (± SD) dropped in the 5- and 10-day group from 39.7 ± 0.7 and 39.8 ± 0.8, respectively, on day 1 to 36.7 ± 0.6 and 36.6 ± 0.4 on days 5–7. Mean WBC count cells (× 1000/mm3) were similar in the 5-day and 10-day treatment groups: 26.7 ± 13.7 versus 26.5 ± 8.4; 10.0 ± 3.1 versus 9.6 ± 3.4 and 11.3 ± 6.7 versus 10.1 ± 3.3, at enrollment, days 5–7 and days 10–14, respectively. The respective mean ANC cells (× 1000/mm3) were also similar between the groups, 20.7 ± 8.4 versus 20.3 ± 6.6; 3.6 ± 2.4 versus 3.1 ± 1.5 and 4.8 ± 2.2 versus 4.0 ± 2.0. The respective mean CRP concentrations (mg/L) were 171.8 ± 170.0 versus 185.3 ± 180.0; 28.0 ± 28.0 versus 16.3 ± 12.0 and 4.3 ± 4.0 versus 4.0 ± 4.0.
The present study demonstrates that a 5-day oral treatment with high-dose amoxicillin (80 mg/kg/d divided to 3 daily doses) is as effective as a 10-day treatment in children 6–59 months of age with nonsevere CAAP. In addition, it is suggested that a 3-day treatment regimen may be associated with unacceptable failure rate in these patients.
Published recommendations for oral treatment duration of nonsevere CAP in the developed world vary, ranging from 3 to 14 days.2,6,10,13 However, these are not based on clinical evidence since there have been no randomized controlled trials in the developed world comparing a short to a long duration of oral antibiotic treatment (5 days or less versus 10 days) for children with nonsevere alveolar pneumonia who are subject to ambulatory treatment.6 Since clinical diagnosis of CAP is made frequently, a shorter antibiotic course, if proven to be adequate, may have a major economic significance, help in reducing antimicrobial resistance,20 improve patient’s compliance and quality of life and reduce adverse events.
In 2 large multicenter, double-blind controlled studies conducted in developing populations, a 3-day oral amoxicillin treatment was deemed as effective as a 5-day treatment for nonsevere childhood pneumonia.5,21 However, these studies may be considered problematic, since they used as enrollment criteria the WHO criteria for treatment of nonsevere pneumonia in the developing world, which are based on clinical findings only. These oversimplified criteria are designed to be used by nonmedical semitrained personnel for treatment purposes and were not meant to qualify for pneumonia studies.2 According to these criteria, patients are diagnosed as having nonsevere pneumonia if cough or difficult breathing are accompanied by fast breathing. Temperature measurements, laboratory findings or chest radiography are not part of these criteria, resulting in selection of children with mainly respiratory viral rather than bacterial diseases, as well as other potentially noninfectious diseases (ie, reactive airways). For this reason, it has been recently recommended to add at least fever as a criterion to the WHO clinical diagnosis of pneumonia in children.22 Furthermore, a recent study showed that for WHO-defined nonsevere pneumonia, placebo was not worse than 3 days of amoxicillin.17 In that study, treatment failure rates judged mainly by tachypnea and hypoxia were high, 7.2% in the amoxicillin arm and 8.3% in the placebo arm. The cumulative treatment failure by day 5 was 13.5% versus 17.6%, respectively (P = 0.09), which prompted the authors to suggest the benefits of antibiotics in children 2–59 months of age who have WHO-defined nonsevere pneumonia are questionable, shedding doubt on the validity of these criteria to discriminate children with respiratory diseases needing antibiotic treatment and those for whom antibiotics are not indicated.
In contrast to the developing countries, criteria for diagnosing childhood pneumonia in developed populations vary widely. Some require evidence of infiltrates by chest radiograph, whereas others require specific respiratory signs or symptoms.2,12,13 We selected children likely to have bacterial pneumonia, and thus included children with all 3 criteria: temperature ≥ 38.5°C, WBC count ≥ 15,000/mm3 and alveolar pneumonia in chest radiography. Although bacterial pneumonia can present in different clinical, laboratory and radiographic forms, it is accepted that the criteria used in our study represent mainly bacterial pneumonia.8,12,23 To the best of our knowledge, this is the first study using predefined criteria to compare duration of treatment in a developed population. We believe that our findings can be extrapolated to other developed populations. Furthermore, the predefined strict criteria, if used in similar studies, will allow comparison between studies, which has not been the case so far.
The most commonly recommended antibiotic for noncomplicated childhood CAAP is amoxicillin.2,10,24 Since most developed countries are routinely using Haemophilus influenzae type b vaccines, there is no need most of the times to use a β-lactamase stable drug (ie, amoxicillin-clavulanate or cephalosporins). In communities with a high rate of penicillin nonsusceptible S. pneumoniae, high-dose amoxicillin is recommended.
Our study has 2 main limitations. First, only a relative small number of patients were enrolled, especially in the 3-day course arm. However, even with this small number, we found significant differences between the 3-day and the 5- or 10-day treatment arms. Although additional studies may be needed, the identical outcome when comparing 5- to 10-day treatments both with regard to failure rate, dynamics of symptoms, signs and laboratory findings make it extremely unlikely that any larger sample size would have detected any differences in outcome between the 2 arms.
Second, we used a very strict definition for pneumonia using a combination of clinical, laboratory and radiographic findings. Although this can, on the 1 hand, be considered as an advantage given the uniformity of the studied subjects, it presents also some disadvantage since the results may not necessarily be applicable to other forms of pneumonia.
We conclude that for children 6–59 months of age with CAAP treated as outpatients, a 5-day high-dose oral amoxicillin treatment regimen is not inferior to a 10-day course. A 3-day treatment regimen may be associated with unacceptable failure rates. Additional studies may be needed to confirm our results.
1. Peltola H, Vuori-Holopainen E, Kallio MJ. Successful shortening from seven to four days of parenteral beta-lactam treatment for common childhood infections: a prospective and randomized study. Int J Infect Dis. 2001;5:3–8
2. WHO. Pocket Book of hospital care for children. Guidelines for the management of common illnesses with limited resources. 2005;4.2 Geneva, Switzerland WHO Press:72–81
3. Atkinson M, Lakhanpaul M, Smyth A, et al. Comparison of oral amoxicillin and intravenous benzyl penicillin for community acquired pneumonia in children (PIVOT trial): a multicentre pragmatic randomised controlled equivalence trial. Thorax. 2007;62:1102–1106
4. Hazir T, Qazi SA, Bin Nisar Y, et al. Comparison of standard versus double dose of amoxicillin in the treatment of non-severe pneumonia in children aged 2-59 months: a multi-centre, double blind, randomised controlled trial in Pakistan. Arch Dis Child. 2007;92:291–297
5. Agarwal G, Awasthi S, Kabra SK, et al. Three day versus five day treatment with amoxicillin for non-severe pneumonia in young children: a multicentre randomised controlled trial. BMJ. 2004;328:791
6. Haider BA, Saeed MA, Bhutta ZA. Short-course versus long-course antibiotic therapy for non-severe community-acquired pneumonia in children aged 2 months to 59 months. Cochrane Database Syst Rev. 2008:CD005976
7. Hazir T, Nisar YB, Abbasi S, et al. Comparison of oral amoxicillin with placebo for the treatment of world health organization-defined nonsevere pneumonia in children aged 2-59 months: a multicenter, double-blind, randomized, placebo-controlled trial in pakistan. Clin Infect Dis. 2011;52:293–300
8. Juvén T, Mertsola J, Toikka P, et al. Clinical profile of serologically diagnosed pneumococcal pneumonia. Pediatr Infect Dis J. 2001;20:1028–1033
9. Principi N, Esposito S. Management of severe community-acquired pneumonia of children in developing and developed countries. Thorax. 2010;66:815–822
10. Harris M, Clark J, Coote N, et al. British Thoracic Society guidelines for the management of community acquired pneumonia in children: update 2011. Thorax. 2011;66:ii1–ii23
11. Bradley JS, Byington CL, Shah SS, et al. The management of community-acquired pneumonia in infants and children older than 3 months of age: clinical practice guidelines by the Pediatric Infectious Diseases Society and the Infectious Diseases Society of America. Clin Infect Dis. 2011;53:e25–e76
12. Esposito S, Cohen R, Domingo JD, et al. Antibiotic therapy for pediatric community-acquired pneumonia: do we know when, what and for how long to treat? Pediatr Infect Dis J. 2012;31:e78–e85
13. McIntosh K. Community-acquired pneumonia in children. N Engl J Med. 2002;346:429–437
14. Triga MG, Syrogiannopoulos GA, Thoma KD, et al. Correlation of leucocyte count and erythrocyte sedimentation rate with the day of illness in presumed bacterial pneumonia of childhood. J Infect. 1998;36:63–66
15. Toikka P, Irjala K, Juvén T, et al. Serum procalcitonin, C-reactive protein and interleukin-6 for distinguishing bacterial and viral pneumonia in children. Pediatr Infect Dis J. 2000;19:598–602
16. The Israel Central Bureau of Statistics. . Statistical abstract of Israel No. 60. 2009;Chapter 2, Table 10 Available at: http://www1.cbs.gov.il/shnaton60/st02_10x.pdf
. Accessed May 2013
17. World Health Organization Pneumonia Vaccine Trial Investigator’s Group. Standardization of interpretation of chest radiographs for the diagnosis of pneumonia in children. 2001 Geneva: World Health Organization WHO/V&B/01.35. Available at: http://www.who.int/vaccine_research/documents/en/pneumonia_children.pdf
. Accessed May 2013
18. Greenberg D, Givon-Lavi N, Newman N, et al. Nasopharyngeal carriage of individual Streptococcus pneumoniae
serotypes during pediatric pneumonia as a means to estimate serotype disease potential. Pediatr Infect Dis J. 2011;30:227–233
19. Rosner B Fundamentals of Biostatistics, 5th Ed. 2000:638–640 Boston, MA: Brooks/Cole, Chapter 13, Section 13.9
20. Greenberg D, Givon-Lavi N, Sharf AZ, et al. The association between antibiotic use in the community and nasopharyngeal carriage of antibiotic-resistant Streptococcus pneumoniae
in Bedouin children. Pediatr Infect Dis J. 2008;27:776–782
21. Pakistan Multicentre Amoxycillin Short Course Therapy (MASCOT) Pneumonia Study Group. . Clinical efficacy of 3 days versus 5 days of oral amoxicillin for treatment of childhood pneumonia: a multicentre double-blind trial. Lancet. 2002;360:835–841
22. Cardoso MR, Nascimento-Carvalho CM, Ferrero F, et al. Adding fever to WHO criteria for diagnosing pneumonia enhances the ability to identify pneumonia cases among wheezing children. Arch Dis Child. 2011;96:58–61
23. Cherian T, Mulholland EK, Carlin JB, et al. Standardized interpretation of paediatric chest radiographs for the diagnosis of pneumonia in epidemiological studies. Bull World Health Organ. 2005;83:353–359
24. Lotha R, Kabra SK, Pandey RM. Antibiotics for community-acquired pneumonia in children. Cochrane Database Syst Rev. 2013;6:CD004874